{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T03:07:23Z","timestamp":1773803243950,"version":"3.50.1"},"reference-count":0,"publisher":"Association for the Advancement of Artificial Intelligence (AAAI)","issue":"29","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["AAAI"],"abstract":"<jats:p>We report a structural mismatch between a data point\u2019s {learnability}\u2014how quickly it improves the loss\u2014and its {forgettability}\u2014how much it anchors the final parameters\u2014an aspect ignored by prior machine unlearning frameworks such as SISA, Fisher-Forget, and influence-based fine-tuning.  \nTo make this gap measurable we introduce Unlearning Gradient Sensitivity (UGS), an influence score computable with a single Hutch++ sketch, and derive the Learnability\u2013Forgettability Divergence (LFD), the Jensen\u2013Shannon distance between the model\u2019s learning and forgetting distributions.  \nWe prove that UGS dispersion decays exponentially only under explicit regularisation and that LFD converges to zero when its weight grows sub-linearly relative to the UGS term.  \nBuilding on these findings, we introduce Dual-Aware Training (DAT)\u2014a lightweight regularization method that reduces variability in how easily data points can be forgotten and aligns learning and forgetting behaviors during training. On CIFAR-10, MNIST, and IMDB, DAT maintains the original model accuracy while cutting forgettability divergence in half and significantly lowering the cost of certified unlearning, showing that it\u2019s effective to make models forgettable from the start.<\/jats:p>","DOI":"10.1609\/aaai.v40i29.39657","type":"journal-article","created":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T01:45:44Z","timestamp":1773798344000},"page":"24718-24726","source":"Crossref","is-referenced-by-count":0,"title":["On the Misalignment Between Data Learnability and Forgettability in Machine Unlearning"],"prefix":"10.1609","volume":"40","author":[{"given":"Zijie","family":"Pan","sequence":"first","affiliation":[]},{"given":"Zuobin","family":"Ying","sequence":"additional","affiliation":[]},{"given":"Yajie","family":"Wang","sequence":"additional","affiliation":[]},{"given":"Wanlei","family":"Zhou","sequence":"additional","affiliation":[]}],"member":"9382","published-online":{"date-parts":[[2026,3,14]]},"container-title":["Proceedings of the AAAI Conference on Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/ojs.aaai.org\/index.php\/AAAI\/article\/download\/39657\/43618","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/ojs.aaai.org\/index.php\/AAAI\/article\/download\/39657\/43618","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T01:45:44Z","timestamp":1773798344000},"score":1,"resource":{"primary":{"URL":"https:\/\/ojs.aaai.org\/index.php\/AAAI\/article\/view\/39657"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,14]]},"references-count":0,"journal-issue":{"issue":"29","published-online":{"date-parts":[[2026,3,17]]}},"URL":"https:\/\/doi.org\/10.1609\/aaai.v40i29.39657","relation":{},"ISSN":["2374-3468","2159-5399"],"issn-type":[{"value":"2374-3468","type":"electronic"},{"value":"2159-5399","type":"print"}],"subject":[],"published":{"date-parts":[[2026,3,14]]}}}